X-ray stereo imaging technique for disparity analysis

Abstract

Presents a simple, fast method for stereo matching of edges on several layers (depths). This method is applied to medical images and to the reverse engineering of printed circuit boards. Differences in images of real-world scenes may be induced by the relative displacement of two cameras. These differences are important because they contain encoded information that can allow a partial reconstruction of the 3D structure of the scene from 2D projections. When such differences occur between two images, the two images are said to have some disparity between them. The determination of disparity is called the correspondence problem. The interpretation of disparity into meaningful statements about the scene, such as specifications of depth, velocity, shape, and so forth, is a field of ongoing research. The primary goal of this paper is to address the correspondence problem using multiple views of a scene with objects at different (discrete) depths using a transparent medium (X-ray imaging). The secondary goal of this paper is to provide a new method of feature (edge) extraction. These features are used for stereo matching. This paper demonstrates the ability to match straight edges and provides an extension to curved edges.<>